U.S. patent number 6,700,885 [Application Number 09/249,762] was granted by the patent office on 2004-03-02 for telecommunications system.
This patent grant is currently assigned to Marconi Communications Limited. Invention is credited to Thomas S Madden, Richard J Proctor.
United States Patent |
6,700,885 |
Proctor , et al. |
March 2, 2004 |
Telecommunications system
Abstract
A telecommunications system comprising one or more
cross-connects and a plurality of telephone exchanges. Two or more
of the telephone exchanges are arranged to communicate with each
other via the one or more cross-connects. An adapter provides the
telephone exchanges with inter-communication via the one or more
cross-connects. The adapter converts traffic between packetized and
non-packetized form.
Inventors: |
Proctor; Richard J (Wimborne,
GB), Madden; Thomas S (Wimborne, GB) |
Assignee: |
Marconi Communications Limited
(Coventry, GB)
|
Family
ID: |
26313123 |
Appl.
No.: |
09/249,762 |
Filed: |
February 16, 1999 |
Foreign Application Priority Data
|
|
|
|
|
Feb 16, 1998 [GB] |
|
|
9803186 |
Apr 1, 1998 [GB] |
|
|
9806862 |
|
Current U.S.
Class: |
370/356;
370/395.1 |
Current CPC
Class: |
H04Q
11/0478 (20130101); H04L 2012/563 (20130101); H04L
2012/5663 (20130101) |
Current International
Class: |
H04Q
11/04 (20060101); H04L 12/56 (20060101); H04L
012/56 () |
Field of
Search: |
;370/258,385,389,395,396,397,400-406,465,466,474,522,535,236,410,495,498,905,907,229,230-235,503,351,352,468,395.1,399,395.21,395.03,395.31,395.4,395.43,395.5,395.52,395.6,395.61,395.64,395.65
;379/219,220,225,226,230-235,242 |
References Cited
[Referenced By]
U.S. Patent Documents
Foreign Patent Documents
|
|
|
|
|
|
|
0 840530 |
|
Oct 1996 |
|
EP |
|
2 274 227 |
|
Jul 1994 |
|
GB |
|
2 316 573 |
|
Feb 1998 |
|
GB |
|
WO 91/15066 |
|
Oct 1991 |
|
WO |
|
Primary Examiner: Ton; Dang
Attorney, Agent or Firm: Kirschstein, et al.
Claims
We claim:
1. A telecommunications system, comprising: a) one or more
cross-connects; b) a plurality of telephone exchanges, wherein two
or more of the telephone exchanges are arranged to communicate with
each other via the one or more cross-connects; c) wherein each of
the two or more of the telephone exchanges comprises routing data
relating to communication with all other exchanges in the
telecommunications system, and wherein the routing data is
partially or wholly enabled; and d) wherein only that part of the
routing data in a particular exchange relating to the communication
between that particular exchange and other exchanges with which
that particular exchange is arranged to communicate via the one or
more cross-connects is enabled.
2. The telecommunications system of claim 1, wherein the
communication via the one or more cross-connects is in the form of
packets.
3. The telecommunications system of claim 2, wherein the
communication via the one or more cross-connects uses asynchronous
transfer mode (ATM).
4. The telecommunications system of claim 3, wherein the
communication uses ATM Virtual Paths (VPs) and/or Virtual Circuits
(VCs).
5. The telecommunications system of claim 4, wherein at least some
of the telephone exchanges arranged to communicate with each other
via the one or more cross-connects are trunk exchanges.
6. The telecommunications system of claim 5, wherein at least some
other ones of the telephone exchanges arranged to communicate with
each other via the one or more cross-connects are local
exchanges.
7. The telecommunications system of claim 6, for handling telephone
calls, wherein all call handling in the system takes place outside
of the one or more cross-connects.
8. The telecommunications system of claim 7, wherein each of the
trunk exchanges has a direct link to each of the one or more
cross-connects.
9. The telecommunications system of claim 8, wherein the
communication between the local exchanges and the trunk exchanges
uses ATM.
Description
BACKGROUND OF THE INVENTION
The present invention relates to telecommunications systems such as
telephone networks comprising a plurality of interconnected
telephone exchanges and communication there-between.
Conventional telephone networks are fully-meshed in that each trunk
exchange has a direct connection to every other trunk exchange.
Traffic levels in telephone networks are increasing, leading to a
continued need to increase the capacity of such networks. To
achieve this, exchanges need to be enlarged and new exchanges
installed. In fully-meshed networks, a significant number of
additional network interconnections are needed with each new
exchange, leading to further increases in the number of ports
needed per exchange. This results in less efficient networks with
costs per unit of traffic increasing.
By using the present invention it is possible to provide a more
efficient and more easily configurable network.
SUMMARY OF THE INVENTION
The present invention provides a telecommunications system
comprising one or more cross-connects and a plurality of telephone
exchanges wherein two or more of the telephone exchanges are
arranged to communicate with each other via the one or more
cross-connects.
In a preferred embodiment the invention provides a
telecommunications system wherein communication via the one or more
cross-connects uses asynchronous transfer mode (ATM).
In a further embodiment the present invention provides a
telecommunications system wherein the communication uses ATM
Virtual Paths (VPs) and/or ATM Virtual Circuits.
The present invention also provides an adapter for providing the
above telephone exchanges with a means of inter-communication of
traffic via the one or more cross-connects wherein the adapter
comprises means for converting the traffic between packetized and
non-packetized form.
BRIEF DESCRIPTION OF THE DRAWINGS
Embodiments of the invention will now be described by way of
example with reference to the drawings in which:
FIG. 1 shows in block diagram form a typical, conventional, large
telecommunications network of the prior art;
FIG. 2 shows the network of FIG. 1 with cross-connects according to
a first embodiment of the invention;
FIG. 3 shows the network of FIG. 2 with cross-connects according to
a second embodiment of the invention;
FIG. 4 shows in block diagram form a typical, conventional, small
telecommunications network of the prior art;
FIG. 5 shows the network of FIG. 4 with cross-connects according to
a third embodiment of the invention;
FIG. 6 shows the network of FIG. 1 partially upgraded with
cross-connects according to the first embodiment of the
invention;
FIG. 7 shows in block diagram form an adapter according to the
present invention;
FIG. 8 shows a possible internal organization of the adapter of
FIG. 7;
FIG. 9 shows in block diagram form the connection of local
exchanges via an AMA; and
FIG. 10 shows in block diagram an alternative connection of local
exchanges via AMAs.
DETAILED DESCRIPTION OF THE INVENTION
Turning to FIG. 1 a typical, conventional, large telephone network
comprises a number of local exchanges L interconnected via trunk
exchanges T. Each local exchange L is connected to two (or more)
trunk exchanges T, which are fully-meshed. In current networks this
fully-meshed trunk interconnect typically uses mainly PDH
transmission.
FIG. 2 shows the typical, large network of FIG. 1 after trunk
upgrade according to the invention. In this embodiment two
cross-connects X are shown, although the number may vary in
practice. Each cross-connect X is connected to every trunk exchange
T. Rather than a large number of low bandwidth connections, the
trunk interconnect now consists of few high bandwidth
connections.
FIG. 3 shows the typical, large network after local upgrade. With
the addition of some simple multiplexers M, all the local exchanges
are directly connected to the cross-connects X. The trunk exchanges
(not shown) may still be used to provide regional interconnect and
interconnection to non-upgraded exchanges and other networks.
FIG. 4 shows a typical, smaller network. The smaller network has
tandems N rather than trunk exchanges, which may have a few
customers directly connected. Each local exchange L is connected to
several (possibly all) tandems N, there is no interconnection of
the tandems.
FIG. 5 shows the typical, smaller network of FIG. 4 after being
upgraded according to the invention. Cross-connects X are
installed, similar to those used to upgrade the large network,
interconnecting the local exchanges L, through multiplexers M. The
old tandems can then be reused, for example, as local exchange and
points of interconnect to other networks.
The cross-connects X can be installed as a trunk network
replacement piecemeal. FIG. 6 shows the large network of FIG. 1
partially upgraded. Here traffic is carried by a mixture of the old
trunk interconnect and new cross-connects X. A number of trunk
exchanges T have connection to the cross-connects X and are able to
communicate with others so connected via the cross-connects.
However not all trunk exchanges T are connected to the
cross-connects X at this stage and the old trunk interconnect is
still used for communication between exchanges not connected to the
cross-connects X and for communication from each such exchange and
each of the exchanges connected to the cross-connects X. The number
of cross-connects X can be upgraded from one to four or more as
further exchanges are connected.
In this way the cross-connects may be extended to eventually
connect all existing trunk exchanges.
It may be desirable to move some of the existing transmission over
to the cross-connects network, to take advantage of the greater
efficiency of the cross-connects X compared with the existing
meshed PDH, and to overcome a shortage of physical ports on the
conventional switches.
Each exchange requires routing data containing information on the
way the exchanges are interconnected within the telecommunications
system. Every time an exchange gets connected to the cross-connects
X, all the other exchanges that have already been connected to the
cross-connects will need to have new routes set up for
communication with the newly connected exchange. Reconfiguring
exchanges in this way every time another exchange is connected via
the cross-connects is expensive. A preferred solution is for
routing data relating to all exchanges to be loaded into each
exchange as it is connected to the cross-connects, but only to
enable data relating to routes to those exchanges that have already
been so connected.
Hence the implementation of the present invention simplifies the
data requirements in exchanges. On installation, every exchange
(trunk or local) can have exactly the same network data loaded.
This data can be pre-prepared for the whole network with the data
for the exchanges connected to the cross-connect being enabled as
the network is enhanced. This data includes the network part of the
digit decode, which is needed to determine the exchange to handle
the call, and the route to be used. The same route number can be
used for all exchanges to get to the same far-end exchange. The
digit decode requirements of the exchanges would grow to support
routing to all local exchanges.
The existing trunk exchanges, could still be used, e.g. as regional
trunk exchanges (DJSUs), allowing reduction or elimination of
sideways routes between local exchanges. If at any time these
regional trunk exchanges are overloaded, the traffic can be handled
in the main trunk (cross-connect) network.
Extension of the cross-connects to directly interconnect local
exchanges may be implemented once there is a sufficiently developed
provision of trunk cross-connect to provide the interworking.
The cross-connects themselves can be relatively simple ATM
connections.
When expanded to connect the local exchanges, enhancement may be
necessary to carry the signalling directly over the ATM
connections. One way of interworking to local exchanges and other
networks is to have simple interworking cross-connects, which are
effectively minimalist exchanges. These could be the existing trunk
exchanges.
The proposed solution supports carrying voice as AAL1 or AAL2 in
configurations to suit different networks at different stages of
evolution.
Telephone calls include traditional hand-set to hand-set calls but
also calls initiated or answered by machine including computers.
Such calls may contain voice and/or data from e.g. modems or
facsimile equipment. Handling of such calls include call set-up,
routing and clear-down. No call handling is required in the central
cross-connects. All the call handling is handled in the PSTN, using
existing protocols with all the existing features.
In a preferred embodiment of the present invention, each
trunk/local exchange has an STM1 (155M) connection to each
cross-connect, carrying one Virtual Path (VP) per destination
exchange. For most local exchanges these 155M connections can be
shared between several exchanges, as detailed below. The VP can
grow or shrink as necessary: the total quantity of traffic is only
limited by the capacity of the transmission to carry it. As traffic
patterns change, VPs may grow or shrink on a dynamic basis: no
configuration is necessary. Furthermore Virtual Circuits (VCs) may
be used in a similar way.
The ATM implementation is capable of carrying some non-voice
traffic but great care is necessary to ensure that it does not
affect the low delay requirements of the trunk voice traffic.
The exchanges communicate using existing SS7 protocols, however
there is an option (that is necessary when it is extended to local
exchanges) to carry the MTP layer 3 over the Signalling ATM
Adaptation Layer (SAAL) on the VP that connects each exchange, thus
eliminating the need for any other signalling hardware.
Further bandwidth savings may be achieved if only active circuits
are packetized. This requires a busy/idle timeslot map to be
carried in an end to end control channel using spare capacity
within the ATM Adaptation Layer. This technique allows the mapping
to be changed dynamically and is set out in detail in patent number
GB 2276518 in the name of GPT Limited.
Optional echo cancellation may be provided but will probably only
be required for the single circuit mode.
Optional signalling transfer point (STP) function to handle SS7
signalling may be provided. This feature will extract redundant
data (Level 2 underframe fill) from the SS7 signalling stream.
Optional modem (fax) detection may be provided. Low speed modems
are not bandwidth efficient. In particular, many connections, e.g.
fax calls, are essentially unidirectional. This optional feature
will detect such calls and convert the modem signals to baseband
data for packetisation into ATM, taking advantage of the fact that
an ATM connection can be asymmetric. Thus further bandwidth savings
can be achieved.
Although this has been described in relation to improve trunk
networks, the present invention could also be deployed into many
other fields for both public and private networks.
Similarly the present invention has application to leased lines
within public networks. Leased lines are often under-utilized,
resulting in idle timeslots being carried across the network. The
use of the present invention, particularly for introducing
communication via ATM VPs will allow the network operator to only
transport the active channels without degrading the service
presented to the end user.
The present invention provides for straightforward enhancement to
support traffic growth wherein the majority of the routing data
needed by most exchanges is identical. This allows operators to add
to and evolve current networks as and when enhancements to
performance or capacity are required.
Additional transmission equipment will be required to bring the
STM1s from each local exchange to each cross-connect. Some ATM
Multiplexer Add drop (AMAs) may be needed to consolidate the
traffic from smaller local exchanges to connect to the
cross-connects, depending on the exact location of exchanges and
where the transmission runs. For interfacing with sources of
non-packetised traffic an adapter is provided, see FIG. 7. The
adapter functions to convert traffic for sending to the
cross-connects into packets and to re-convert packets received from
the cross-connects to un-packetised form. In one embodiment the
adapter provides interworking between a PSTN trunk (G.703/G.704)
and an ATM domain to allow ATM to be used to provide an efficient
and flexible interconnect between exchanges by using variable
bandwidth (ATM) routes.
For interfacing with STM networks, the adapter provides
interworking between ATM and STM, with preferably one adapter for
each route via the cross-connects. These adapters perform
interworking between 2M bit/s PCM traffic and ATM and also provide
a means to connect signalling directly with each of the other local
exchanges. These boxes communicate with each of the processor
clusters as most of the signalling traffic passes through them.
The adapter box can preferably be programmed for different types of
adaptation to allow for various ATM adaptation methods, and any
other special handling.
We now consider the adapter construction in more detail. Referring
to FIG. 8, each adapter unit comprises a housing with a PSU, a
common card and up to 4 PCM cards. Each PCM card could handle up to
16 PCMs, the common card has an STM1 interface, an Ethernet
interface and a controller for the PCM cards. Each exchange will
need 4 adapters in this embodiment and these could preferably all
be mounted within one shelf of the exchange. If provision is made
for T1 rates (as well as E1 rates) backplane provision for a fifth
PCM card may be desirable.
The cross-connects operate under PSTN control and provide no
control processing themselves. Advantageously a small number of
cross links are provided between the main cross-connects, to
provide some flexibility after major transmission network outages.
These would only be brought into use when a pair of local exchanges
have no common accessible trunk cross-connects (a very rare
occurrence).
The routing function selects the single VC per VP that handles the
signalling and routes that VC to the SAAL handler and thence to the
processor interconnect. The rest of the traffic is routed to the
appropriate PCM/Circuit handler. The SAAL handler might also, if
required, handle some proprietary maintenance channels between the
adapter units, for example to prevent potential overload when
concentration of small local exchanges is provided.
The processor interconnect handles the signalling traffic, the
mapping messages from CPS and general management. A suitable
interface to use here is Ethernet as it allows the interface to be
fairly open for use on other platforms if necessary.
The hardware of the PCM termination could support both E1 and T1 or
be dedicated to E1 data rates.
With known networks, it would be a waste of resources to connect
all the small local exchanges directly to each of the four ATM
cross-connects via full STM1 connections. By using AMAs to provide
simple ATM concentration, a number of small local exchanges can be
grouped together to share a single STM1 link to the ATM
cross-connects.
A typical large network might require concentration of roughly 6
local exchanges onto one STM1. This concentration could be provided
in two ways: at common points star connected local exchange could
share a single ATM Multiplexer Add drop (AMA), or a small ATM
switch (see FIG. 9); alternatively a number of local exchanges
could be chained or connected in a ring by using a small number of
smaller AMAs (see FIG. 10).
It is necessary to limit the traffic that such concentrated local
exchanges would load onto the shared STM1 connection. This can be
achieved whilst avoiding complex signalling as long as the network
configuration limits the total number of PCMs connected to the
adapter units in each of the local exchanges to 64. This could be
ensured by the adapter units passing suitable messages between
themselves.
This type of trunk network has the potential capacity to grow to be
significantly larger than today's, without needing any significant
changes. As the traffic grows, fewer local exchanges can be
concentrated into multiplexers and more will need a direct
connection to the ATM cross-connects. This leads to a need for
larger ATM switch cross-connects and additional transmission. In
the extreme, each local exchange could require many whole STM1s
allocated to it without any concentration to the central ATM
cross-connects. The number or size of the central ATM
cross-connects needed to handle the increased traffic would
increase accordingly.
The adapters also exchange status information with each other to
ensure that each is operational and to exchange and pass on
traffic-blocking information. Periodically and when necessary (for
example because of a status change) each adapter should send a cell
to each far end adapter. This cell contains three elements: 1. The
adapter receiver status (i.e. whether the adapter is happy with
what it is receiving from the far end adapters); 2. A hold
indication that, if set, should cause the far end adapter to stop
the flow of new traffic to that adapter; 3. Status information on
the PCMs to the exchange (i.e. reflecting which PCM/circuits can be
used).
When an adapter receives a check cell, it reacts as follows: 1. If
the far is not happy, it should force clear all calls and hold that
route busy (but still send its own check cells); 2. If the far end
requests a hold, send a hold on the Q.50 to the associated
exchange; 3. If the far end ceases to request a hold clear the
condition (the adapter may delay reacting to a change of state to
prevent the network oscillating); 4. If the far end has PCM
failures, reflect these on to the alarm indication signal (AIS) of
its transmitted PCMs.
The adapter also packetizes 64 kbit/s circuits into ATM cells in
three modes:
single circuit AAL1 (6 ms delay);
multiple circuit AAL1 (about 250 microseconds delay for 20
circuits); and
multiple circuit AAL2 (about 250 microseconds delay for 20
circuits).
The 64K channels could be carried in many different ways, using
AAL1, DBCES, AAL2 and in proprietary ways. They could be packaged
as single channels or as N channels. The choice is based on a
trade-off between delay and inefficiency due to unbalanced traffic.
The adapter of the present invention can support all of these ways
of carrying the 64K channels and allows changes as the network
design evolves, for example starting as N channel packed, then
moving to single channel packed, or perhaps AAL2.
To minimize the delay, at the expense of network utilization, N
channels could be carried in one ATM Virtual Circuit, e.g. a whole
PCM. This is most appropriate for trunk interconnects where the
route sizes are large. Some compression may be achieved by removing
unused channels. Capacity freed in this way could be used to
support lower priority data traffic.
With this method of interconnecting local exchanges, the network
efficiency is low with partially filled routes using a whole cell
for one call and (N-1) other channels being wasted.
The ATM Forum specification for dynamic bandwidth circuit emulation
service (DBCES) compresses out unused channels on N channel
packetized ATM.
To maximize network utilization and simplify the signalling, at the
expense of delay, the 64K channels could be carried in individual
VCs. The main problem of this is delay as there would be nearly 6
ms packetization-delay to fill a cell. This would however result in
a very simple and easy to understand network, both for traffic and
data management.
Partially filled cells could be used to minimize the delay. It may
only be necessary to partially fill a few cells for the longest
routes, the rest being unaffected by the packetization delay. For
deployment to the local exchanges, this method avoids very large
wastage of network bandwidth resulting from the route sizes and
traffic imbalance.
AAL2 has been designed to carry low rate, short length packets in
delay sensitive applications in a bandwidth efficient way. The AAL2
multiplexes many separate low speed streams such as voice, data and
signalling together on a single ATM VC. AAL2 has been standardized
in ITU-T I.363.2. AAL2 is more efficient than sub-loaded ATM
cells.
AAL2 could be used by the adapter to support trunk interconnect in
a very similar way to the N channel packet defined above, though
with increased delay. It would be very appropriate for environments
where transcoding is handled and direct compression.
Voice traffic is given a high priority. This ensures that the delay
variation would be well constrained, thus limiting the amount of
additional buffering required to remove the effect of cell delay
variation at the receiver.
The provision of the ATM infrastructure allows a simple migration
of the signalling transport. Signalling messages are carried
between the exchanges by encapsulating MTP L3 over the signalling
AAL (SAAL) and then sending these between the exchanges. This would
be faster and involve less equipment than conventional signalling
methods.
Advantageously, all the signalling transport could be handled by
the ATM network . Signalling using existing MTP L2 into the Digital
Switching Subsystem (DSS) could be advantageously handled by the
processing power of the adapter unit directly.
The Control protocol between the call processing and the adapter
units is implemented as an application programming interface (API).
This API covers the setting up and breaking down of switching
associations, and optionally management (e.g. failure and
configuration management).
Management of the adapter includes managing the equipment,
configuring it and handling any appropriate statistics. Management
is preferably achieved through the exchange or, when the adapter is
being used as a standalone device, a Simple Network Management
Protocol (SNMP) interface would be appropriate for it to be managed
by third party management systems. Alternatively a standalone
adapter could be managed through the Equipment Management
Operations System (EMOS) particularly when they are closely
associated with transmission. Thus, depending on where and how the
adapter is being used, different management solutions are
necessary. Within System X, the adapter could be considered as part
of MTS. By treating it as part of MTS it will minimize the affect
of the adapter on the rest of System X.
When the adapter is being used as standalone equipment, a SNMP
management interface over the Ethernet control port will be
appropriate. This will be controlled by a Managed Information Base
(MIB) which should be published. The data model necessary for these
applications may differ significantly depending upon the type of
exchange used. Flexibility is essential to allow for the addition
of functionality such as transcoding, packing algorithms and data
rate change. When the adapter is deployed along with SDH
transmission, it may be appropriate to manage it through EMOS and
the rest of the SDH. This could be through the Ethernet port,
through the DCC of the STM1, or through an ATM VC.
In principle all the traffic could be handled by a single
cross-connect, but this is likely to be unacceptable in practice
for availability reasons.
There will preferably be at least two cross-connects to allow for
catastrophic failure. However even with two it will be necessary to
have a very high level of redundancy with alternative routing for
the transmission to these cross-connects. With four cross-connects
there is a good level of built-in network redundancy, without the
cost being too high. Each should preferably be dimensioned such
that any two could handle the normal peak traffic loads of the
entire network.
For maximum availability suitably diversely routed transmission
should be arranged to the cross-connects with each cross-connect on
an independent site.
The basic architecture would work with any of the following
combinations:
Two cross-connects connected over two diverse routes and using
transmission protection; three cross-connects over two diverse
routes using transmission protection or three cross-connects with
three diverse routes without transmission protection. Four
cross-connects over two diverse routes without using transmission
protection; four cross-connects over two diverse routes using
transmission protection and six cross-connects connected via two to
three diverse routes without transmission protection. Obviously the
numbers above could be varied by the skilled worker depending on
the requirements concerning resilience to failure and cost.
* * * * *